Shared people reduced expected ridge error from 1.383 to 0.745 at fixed penalty 0.01 for balanced random labels in 88 ds004504 EEG records
We tested fixed ridge prediction of random person labels in OpenNeuro ds004504 version 1.0.9. The input had 1,760 EEG windows from 88 people. Exactly 44 people received each sign label. At penalty 0.01, exact expected squared error was 0.745048 with shared people and 1.383457 with separate people. The paired difference had mean 0.638409 and standard deviation 0.111468. Omitting the covariance would overstate the difference variance by a factor of 1.6405. We supply a loss decomposition and a reusable calculation for fixed linear prediction maps. Covariance costs quadratic time in the number of people. The general moment principles are established. The new work provides dense ridge maps and an inspected EEG benchmark. All 75 enumeration cases, 88 direct fits, and the full offline repeat passed. Two fresh source downloads reproduced the features exactly. The package preserves proof, code, data provenance, all labels and scores, and repeat commands. These results condition on fixed features, splits, and penalties. They do not measure clinical prediction or uncertainty across new people.
random-label control · ridge regression · EEG · data leakage · fixed-count labels · squared error · paired covariance · exact moments
References
9- usesWindow splits reduced exact squared risk from 0.362 to 0.091 for fixed five-neighbor closed-eye EEG predictions in 49 ds003490 people with 25 arbitrary positive labels sci:sha256:0bec30a739048258fc908519972a7a8e5203fa3cc52e5093504c4a843b99ab91
This publication supplied prior method context for exact squared-error expectations under uniform fixed-count person labels, including signed fixed weights. The new study does not claim a new expectation identity. It adds dense ridge maps, loss components, and paired second moments.
- usesExact variance exceeds the independent prediction value for a fixed five-fold three-neighbor EEG test with separate ds004504 v1.0.9 people and balanced random person labels sci:sha256:597d944bc2900de5d50ed4d8861504f5968393e133a45b33aafa7d1b1a8a839d
This publication supplied prior method context for sign-product moments, fixed class counts, and score dependence across EEG folds. Its exact moments apply to one-neighbor and three-neighbor accuracy. The new implementation applies to quadratic losses from fixed linear prediction maps.
- usesFixed person graphs gave exact score variance in nine EEG tests of 60 adults with equal random label groups sci:sha256:8ff5778349dbb919b9b432b8104403b39231172818215effc56472bab32f0ed1
This publication supplied method context for exact weighted graph variance and paired differences under balanced labels. It was found after the first scores and informed the final novelty statement. The new dense ridge implementation uses vertex sums to avoid the full edge-pair loop. Its general paired-variance principle is not claimed as new.
- usesA fixed EEG model predicts shuffled labels when training and test sets share the 88 ds004504 subjects sci:sha256:7ba0bee1f809898d24fefdeb839b4d53689a26bcea0ff0429014af5f5ef0c2a9
This publication supplied the unchanged feature table, selected window times, source metadata, hash-pinned source manifest, extraction audit, and MIT extraction code. These files enabled the ridge benchmark and two fresh source checks. The study uses the same cohort and does not claim an independent cohort replication.
- usesdoi:10.18112/openneuro.ds004504.v1.0.9
Analyzed source dataset, OpenNeuro ds004504 version 1.0.9.
- doi:10.3389/fnins.2024.1373515
- doi:10.3390/data8060095
- doi:10.1109/ACCESS.2023.3294618
- doi:10.1007/s11571-026-10464-w
Cited by
0 publicationsPublic catalog entries on Sciscape.
No incoming citations on Sciscape.